CONTENTS
Preface = ⅶ
1 Overview = 1
1.1 What are expert systems? = 1
1.2 How expertise is acquired : the debug cycle = 3
1.3 Induction = 4
1.4 Inductive knowledge acquisition environment = 8
1.5 Applications = 8
1.6 Duce applications = 10
1.7 Summary = 10
2 Inductive inference = 13
2.1 Generalisation = 13
2.2 Examples and rules = 14
2.3 Criteria for inductively generated results = 16
2.4 Languages involved in inductive inference = 18
2.5 Classification learning = 19
2.6 Finite-stage automata and strategy learning = 24
2.7 Induction of finite-state automata = 30
2.8 Conclusion = 31
3 RuleMaster = 33
3.1 Some issuesin knowledge engineering = 33
3.2 RuleMaster = 34
3.3 Knowledge acquisition = 35
3.4 Types of expert systems supported = 38
3.5 Radial = 39
3.6 Individual Radial modules = 45
3.7 Operator definitions = 50
3.8 Explanation = 51
3.9 The Rulemaker code generator = 53
3.10 External information sources = 60
3.11 Conclusion = 61
4 Robotic applications = 63
4.1 Introduction = 63
4.2 The problem : buildig a five-block arch = 63
4.3 The action arch = 65
4.4 The action onto = 66
4.5 The action clear = 68
4.6 A session = 68
4.7 GENARCH = 71
4.8 Conclusion = 81
5 Expert systems applications = 83
5.1 Introduction = 83
5.2 SHUTTLE = 83
5.3 WILLARD = 86
5.4 EARL = 91
6 Grammatical induction theory = 95
6.1 Introduction = 95
6.2 Language identification = 97
6.3 Mixed positive/negative presentations = 98
6.4 Positive samples = 98
6.5 An efficient new algorithm = 109
6.6 k-contextual languages = 115
6.7 Use of semantic information = 120
7 sequence induction applications = 127
7.1 Introduction = 127
7.2 A simple grammar = 128
7.3 1-bit binary adder = 130
7.4 Traffic light controller = 131
7.5 Reverse motor problem = 134
7.6 Algebra Problem = 136
7.7 Hanging pictures in a room = 140
7.8 Conclusion = 143
8 Chess strategies = 145
8.1 Introduction = 145
8.2 The problem : KBBKN = 147
8.3 Conclusion = 152
9 Duce = 153
9.1 Introduction = 153
9.2 Background = 154
9.3 Transformation-based induction = 154
9.4 Operators = 155
9.5 The search algorithm = 158
9.6 Animal taxonomy = 159
9.7 Even-parity = 162
9.8 Recreation of the KPa7KR structure = 165
9.9 Neuropsychology application = 168
9.10 Conclusion = 169
Appendices
A ACLS, ID3 and CLS = 173
A.1 The entropy function = 174
B Definitions = 177
C Heuristics used in the Iiterature = 181
C.1 Biermann and Feldman's k-tail predicate = 181
C.2 Levine's heuristic = 182
C.3 Miclet's algorithm = 182
C.4 Angluin's algorithm = 182
D Proofs = 185
E Example move sequences = 195
E.1 Actions = 195
E.2 Attributes = 195
E.3 Black plays optimally = 196
E.4 Black plays badly = 198
F Results of sequence induction = 201
F.1 Actions = 201
F.2 Attributes = 201
F.3 State machine = 202
G. Automata after ACLS induction = 203
G.1 Actions = 203
G.2 Attributes = 203
G.3 State description = 204
H KBBKN Rulemaker induction file = 205
I KBBKN Radial code = 207
Bibliography = 209
Index = 215
